Jump to content

Hit detection, All that really matters


140 replies to this topic

#121 Erhardt

    Member

  • PipPip
  • Knight Errant
  • Knight Errant
  • 43 posts

Posted 05 November 2011 - 05:09 PM

View PostCaveMan, on 05 November 2011 - 03:35 PM, said:


Incorrect. The magnetic coils exert a force on the projectile, but they exert an equal force on the 'Mech. It's simple physics. The law of conservation of momentum says that the total momentum of a system is constant, meaning the projectile and launcher experience equal impulse. Otherwise things like ion drives wouldn't function. Of course the 'Mech can still have recoil dampers, like having the whole gun floating in oil, but there are practical limits to what these systems can absorb. You're going to feel some recoil no matter what.

I think you're right, but it doesn't seem, to my mind, anyway, that it'd be anywhere near the recoil you'd get from a shell explosively driven out the gun.

View PostCaveMan, on 05 November 2011 - 03:35 PM, said:

Not entirely true about fighter jets. There's no risk of burning up at Mach 3. The reason titanium and other high-temperature alloys are used is because aerodynamic heating softens the metal. At mach 3 leading edges only reach about 1500F. At Mach 70, you're correct, there would be significant ablation of the projectile due to friction. However at 24,000 m/s it would cross several kilometers before enough mass was lost to render it ineffective. Also, nickel-iron has a higher melting point than titanium :)

I guess it was weight consideration then when Titanium was chosen for the high-mach jets. *shrug*\

View PostCaveMan, on 05 November 2011 - 03:35 PM, said:

I still don't agree about hitscan weapons. Even if we can completely ignore ballistic drop and windage effects (which the tabletop rules say we can't, as a mere 660m carries a huge +4 range modifier, indicating significant aiming is required, otherwise you could eyeball a shot like that on a 12m target), hitscan gives credit to the guy with the fastest mouse polling and the shortest ping. At least some ballistic tracking is needed even if it's just a formality.

Eh. I want the game to be simmy at an operational level. If they want to get all ARMA 2 on weapons-delivery, that's fine, but I'll be every bit as happy if that aspect isn't gonzo-hardcore.

#122 Neozero

    Member

  • PipPipPipPipPip
  • Rage
  • Rage
  • 136 posts

Posted 05 November 2011 - 05:10 PM

View PostRIFT, on 02 November 2011 - 07:26 AM, said:


World of Tanks dose this now, I dont see why it cant be done in this game.

World of tanks does not have subsections of there tanks aside from Turret/Hull. mechs are a bit more complex then tanks and will feature both external and internal armor structures. The standard model for CBT will translate just fine to PC MW as long as they include internals and make its o legged mechs are not automatically dead. Mechs die when they have no gyros or no engines and thats it.

#123 CaveMan

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,127 posts
  • LocationIn a leather flying cap and goggles

Posted 05 November 2011 - 05:28 PM

View PostNeozero, on 05 November 2011 - 05:10 PM, said:

World of tanks does not have subsections of there tanks aside from Turret/Hull. mechs are a bit more complex then tanks and will feature both external and internal armor structures. The standard model for CBT will translate just fine to PC MW as long as they include internals and make its o legged mechs are not automatically dead. Mechs die when they have no gyros or no engines and thats it.


This is true.

A Tank is, fundamentally, a metal box with some very delicate machinery and squishy meat inside. The whole goal of tank design is to make sure nothing ever penetrates that box, because if it does, even the slightest hit getting through will kill your crew or set off the ammo. There's a reason HESH rounds are so nasty: they don't even have to break the armor, they just spall bits of it off inside the cabin and they bounce around like bullets.

Mech design is a totally different affair. Armor is there mainly to protect internal components, of course, but those components are widely spaced and generally not part of the same open volume. They tend to be quite well protected from most angles, and the loss of one component doesn't disable the whole 'Mech. Modeling the complex interaction between armor, internal structure, and components, and the effects of angles and ricochets inside a 'Mech would, frankly, require some serious simulation--on a supercomputer level, most likely, to do it in real-time. Not to mention each chassis would need hundreds of hours of modeling to properly model all the components. All that stuff takes time and money, resources that a F2P game isnt going to have starting out.

#124 Vaneshi SnowCrash

    Member

  • PipPipPipPipPipPip
  • 339 posts

Posted 05 November 2011 - 07:17 PM

View PostMookieRah, on 02 November 2011 - 08:02 AM, said:

I don't see why there is so much hate for good physics in game.


Most people aren't but most people will understand that this is an MMO, your local machine is doing very little of the work. The backend cluster does it because you NEVER trust the client (anyone saying you should trust the client likes aimbots). Now even if my worst nightmare comes true and this is a lobby based map rotating FPS you have to think: How many matches are going on now, with how many people with how many shots being fired?

Each one of those shots will need to go in to the physics processing system and the results sent to your client for display. That is computationally expensive, which translates in to needing a much larger cluster than you would with a simpler system that'd please 95% of the people wanting to play your game. Blades ain't cheap, neither are the bills associated with the care and feeding of a cluster.

#125 Tierloc

    Member

  • PipPipPipPipPipPip
  • 231 posts
  • LocationWAR_Homeworld

Posted 05 November 2011 - 08:32 PM

View PostCavadus, on 01 November 2011 - 06:50 PM, said:

It would behoove them to actually, like, y'know... make a good PC game first and worry about some board game rules meant to be calculated by hand written thirty years ago later.

This needs to be copy/pasted in another 300+ threads.


NVidia has Physx, which may be an answer for this.
http://www.nvidia.co.../physx_faq.html

#126 zverofaust

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,093 posts
  • LocationMontreal

Posted 05 November 2011 - 10:09 PM

View PostErhardt, on 01 November 2011 - 02:53 PM, said:

I wouldn't hold my breath for that kind of damage model if I were you. Hit detection will be paramount, but burning cycles of code on computing mass and velocity of a Gauss round round vs. angle of attack and density of armor then computing damage applied from there probably isn't in the cards when we're already not getting simple straightforward physical attacks or AI opponents. I'm sure it'll be more along the lines of the traditional "did it hit? yep. apply 15 pts of damage. done"


Not necessarily. World of Tanks seems to use at least a somewhat advanced model that takes armour thickness, angle of impact, speed and kinetic energy into consideration. Sometimes rounds with just soundly thud against armour, other times richochet off, other times penetrate.

#127 Alizabeth Aijou

    Member

  • PipPipPipPipPipPipPip
  • 877 posts

Posted 06 November 2011 - 02:27 AM

View PostNeozero, on 05 November 2011 - 05:10 PM, said:

World of tanks does not have subsections of there tanks aside from Turret/Hull.

Sponsons!
Don't forget the smegging sponsons, dammit!

Quote

Mech design is a totally different affair. Armor is there mainly to protect internal components, of course, but those components are widely spaced and generally not part of the same open volume.

Pretty much. Here's a cutaway showing the internals of a Timber Wolf (Mad Cat)
Posted Image
(Source: TechManual CGL35002)

Quote

your local machine is doing very little of the work

Oh, it'll have to work darn hard to render the entire internal structure of a single 'Mech, though.
Imagine an FPS game that renders all the muscles, tendrons, organs, bones, etc. of a human, and for the fun of it lets have a realistic blood circulation on that as well. Good luck getting your comp to render more than a pair without becoming a slideshow.

#128 Redraider

    Member

  • PipPip
  • 23 posts
  • LocationSeattle

Posted 07 November 2011 - 11:27 AM

View PostCaveMan, on 05 November 2011 - 03:57 PM, said:

I like the suggestion upthread (or was it in a different thread?) about keeping weapon mounts in different body sections parallel. Got 6 lasers in an arm? Good luck pinpointing that group shot into the enemy's cockpit. It would essentially require precision shooters to use chain fire and individually aim each weapon in sequence--much harder to do than with converging guns.


Or a method similar to WWII fighters in which the weapons had a convergence point set by the pilot at a specific, unchangeable distance. Then fighting at a certain distance has distinct advantages.

#129 Gaius Cavadus

    Member

  • PipPipPipPipPipPip
  • The Spear
  • The Spear
  • 404 posts
  • LocationNova Roma, Alphard

Posted 07 November 2011 - 11:39 AM

View PostNeozero, on 05 November 2011 - 05:10 PM, said:

World of tanks does not have subsections of there tanks aside from Turret/Hull.


This isn't true at all. Every surface of the tank has an armor value assigned to it and various parts of the tank can significantly deviate from the listed values in their info screen.

A good example is the IS-4's frontal armor. It's listed as 160mm thick but not every square inch of the front of the tank actually has that listed thickness. For instance, the driver's hatch on the front of the tank directly beneath the turret ring only has 80-100mm thickness.

All of their tanks are assigned armor thickness in this way. There many weak spots, each surface has it's own armor thickness value, and all of the modules are internally modeled with their own hitpoint pools. It's quite complex when you get down to it.

#130 Redraider

    Member

  • PipPip
  • 23 posts
  • LocationSeattle

Posted 07 November 2011 - 12:36 PM

I had previously stated that a multiple hitbox model could work depending upon the mechanics that determine damage along side it.

The perfect pinpoint aiming system of the previous games needs to go. No weapon ever designed fires into the exact same location every time and nothing about the classic system or real life lends itself to the dead aim of previous games. That isn't to say I propose missing the broadside of a barn but we can't reasonably simulate the forces at play on a battlefield so some reticule bloom is necessary. Then I suggest that the hitbox's be components of the greater whole instead of some facade over an "I win" button.

Using an Atlas as an example, the thigh isn't a solid block, it's a component made up of several sub assemblies. I would hope that this thigh was divided into six specific pieces. Three external armor sections and three internal structure/systems sections. The armor has two roles, it absorbs damage prior to being destroyed and degrades damage after its destroyed. Structure also degrades damage after destroyed. The damage degradation represents the "junk" leftover in the location absorbing damage prior to it impacting the structure or armor in the path of the weapon. This would mean blowing a leg or arm off would be considerably more difficult than previous iterations of the game where anywhere on the leg was acceptable. The Mech's design would mean it had weak points and strong points. A thigh could have four sections instead of three while the knee only has two but would be a much smaller target that is constantly moving around. Then taking out a leg would be a matter of taking out the entire structural assembly of that leg instead of the underlying structure being one all encompassing component. It should make leg sniping much more difficult than it currently is without straying too far from the classic model, add a bit of play style balance to the game, and provide a bit of realism at the same time.

This is just my opinion however and I doubt it will come to fruition. I imagine the same model that is used in previous games will be used and I will just hop in my Dire Wolf or Uzial and one shot or snipe people to death.

Edited by Redraider, 07 November 2011 - 12:41 PM.


#131 Kaemon

    Member

  • PipPipPipPipPipPipPipPip
  • 1,924 posts
  • LocationMN

Posted 07 November 2011 - 12:53 PM

View PostUncleKulikov, on 01 November 2011 - 08:16 PM, said:

Client Side hit detection would help a lot of latency issues, and could allow what you see on your screen, to happen to them.
anything client side is open for hacking, you really want to give people the ability to mod the hit detection?

#132 Kaemon

    Member

  • PipPipPipPipPipPipPipPip
  • 1,924 posts
  • LocationMN

Posted 07 November 2011 - 12:56 PM

View PostCavadus, on 07 November 2011 - 11:39 AM, said:


This isn't true at all.  Every surface of the tank has an armor value assigned to it and various parts of the tank can significantly deviate from the listed values in their info screen.

A good example is the IS-4's frontal armor.  It's listed as 160mm thick but not every square inch of the front of the tank actually has that listed thickness.  For instance, the driver's hatch on the front of the tank directly beneath the turret ring only has 80-100mm thickness.

All of their tanks are assigned armor thickness in this way.  There many weak spots, each surface has it's own armor thickness value, and all of the modules are internally modeled with their own hitpoint pools.  It's quite complex when you get down to it.
If you download the skin vulnerability pack for WoT you'll see alot of weak points in certain areas (hatches, ports, etc.) which make sense and are supposedly somewhat accurate historically.Similarly I hope that some of the mechs have weakness in joint and locations that it makes sense, and that someone creates a client side skin that shows these (so I can actually do some damage to people who are better players than I am).

#133 Cake Bandit

    Member

  • PipPipPipPipPipPipPip
  • Veteran Founder
  • Veteran Founder
  • 500 posts
  • Facebook: Link
  • LocationHipsterland, USA

Posted 07 November 2011 - 01:22 PM

I like the idea of having all these little armor variances. There's a whole lot of talk about what they're "probably" going to do and to that I say pish posh. They've got nothing but opportunities ahead of them, and I think that if they can manage it, it should be done. It would be fine to add a second or two of delay to ballistic weapons while the onboard targeting computers take into account range for aiming purposes. If you were a good judge you could do it manually before the computer makes it happen. Yeah, it would pose some difficulty for weapons of multiple ranges in the same firing group, but that would just be part of the sacrifice for making snap shots like that. Believe it or not, these are giant tanks, I think a little bit of tank-scale gun physics is entirely reasonable if they want to implement it.

At the very least we should stop and consider how it might affect gameplay before we just throw up our hands and say "But that'll be haaaarrrrrrd"

#134 TheRulesLawyer

    Member

  • PipPipPipPipPipPipPipPip
  • 1,415 posts
  • LocationChicagoland

Posted 07 November 2011 - 01:59 PM

If anything at all has become clear from this discussion, it should be that battletech and realistic physics don't belong in the same sentence.

#135 Zakatak

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,673 posts
  • LocationCanadastan

Posted 07 November 2011 - 03:22 PM

So you want a complex damage system?

Hitting the the middle of the CT with a high velocity, high density weapon (light gauss) will pierce the fusion chamber and create big boom?
Hitting the lower cockpit will cause electronics to be fried, and hitting the upper cockpit will decapitate the pilot?
Hitting the joints of the legs will disconnect the myomer and topple the mech, while hitting the thigh accomplishes little?

The problem with this is what I saw in Crysis (that had many different hit multipliers). This didn't make combat more tactical or aim-orientated, it just made it all about "lucky shots".

#136 Alizabeth Aijou

    Member

  • PipPipPipPipPipPipPip
  • 877 posts

Posted 08 November 2011 - 06:43 AM

Quote

Hitting the the middle of the CT with a high velocity, high density weapon (light gauss) will pierce the fusion chamber and create big boom?

Doesn't work that way, only in Stackpole novels it does.

Quote

Hitting the lower cockpit will cause electronics to be fried, and hitting the upper cockpit will decapitate the pilot?

2 Sensor crits, 2 life-support crits, 1 CPit crit.
Crit the CPit and you've pretty much disabled the 'mech, the others simply make it less effective.

#137 Glare

    Member

  • PipPipPipPipPip
  • FP Veteran - Beta 1
  • 192 posts
  • LocationAtreus

Posted 08 November 2011 - 10:34 AM

The introductory fiction piece for the BattleMech proper in the TechManual rulebook pretty specifically and completely covers fusion engine detonation.

Summed up: Unless you make it your number one priority to blow yourself up, it's not going to happen. It just isn't. And doing that doesn't really make a ton of sense unless it's for a predetermined strategic or tactically vital reason, which doesn't come up that often.

#138 GreenHell

    Member

  • PipPipPipPipPipPipPip
  • 543 posts
  • LocationGrandmas House

Posted 08 November 2011 - 01:06 PM

View PostTierloc, on 05 November 2011 - 08:32 PM, said:

This needs to be copy/pasted in another 300+ threads.


NVidia has Physx, which may be an answer for this.
http://www.nvidia.co.../physx_faq.html



I actually have a dedicated PhysX processor in my rig. Not many games make use of that hardware. I'd be in love if they made use of it! :)

#139 UncleKulikov

    Member

  • PipPipPipPipPipPipPip
  • Bridesmaid
  • 752 posts

Posted 08 November 2011 - 01:13 PM

View PostRedraider, on 07 November 2011 - 11:27 AM, said:


Or a method similar to WWII fighters in which the weapons had a convergence point set by the pilot at a specific, unchangeable distance. Then fighting at a certain distance has distinct advantages.

That would be bad, since mechs can carry a mix of variably ranged weaponry and movement is much less free. Having all of your weapons matched to your shortest ranged weapon (lets say machine guns) and then your LRMs wouldn't work at all.

Good idea, bringing in other simulation games, but it just wouldn't work well.

#140 Redraider

    Member

  • PipPip
  • 23 posts
  • LocationSeattle

Posted 08 November 2011 - 02:14 PM

View PostUncleKulikov, on 08 November 2011 - 01:13 PM, said:

That would be bad, since mechs can carry a mix of variably ranged weaponry and movement is much less free. Having all of your weapons matched to your shortest ranged weapon (lets say machine guns) and then your LRMs wouldn't work at all.

Good idea, bringing in other simulation games, but it just wouldn't work well.


Bringing a mixed bag into a battle is generally the reason people fail at PVP. This isn't going to be the board game where a lucky dice role can determine the outcome of the battle.





12 user(s) are reading this topic

0 members, 12 guests, 0 anonymous users